Methods

The design and results of the FUSION II trial have been previously reported. FUSION II was a prospective, randomized, parallel, multicenter, double-blind, placebo-controlled trial in patients with advanced HF receiving appropriate standard background therapy. Patients with 2 recent HF hospitalizations, left ventricular ejection fractions <40%, and New York Heart Association class IV symptoms or New York Heart Association class III symptoms with creatinine clearance <60 ml/min were randomly assigned to receive nesiritide or matching placebo once or twice weekly for 12 weeks. All patients were treated to optimal goals with evidence-based medical or device therapy facilitated by careful disease management during the study. The primary end point was time to all-cause death or cardiovascular or renal hospitalization at 12 weeks. Enrollment began in April 2004 and ended in June 2006; follow-up ended in December 2006. The trial included patients from North America, Central America, South America, Australia, Russia, India, and the Pacific Rim. The protocol was reviewed and approved by the institutional review board at each participating institution. All participants provided written voluntary informed consent.

The CEC was organized through the Duke Clinical Research Institute. The CEC developed a charter, defined clinical end point events consistent with the protocol, developed algorithms to identify patients with suspected end point events from data in the case report forms (CRFs), worked with monitoring groups to collect supplemental medical records for event review, and adjudicated suspected clinical end point events. Events were adjudicated by 5 cardiologists blinded to treatment assignment. Events resulting in death or hospitalization were adjudicated according to prespecified criteria. The CEC classified deaths and hospitalizations as cardiac, renal, cardiorenal, other or noncardiovascular, or unknown. The cardiac category included HF causes. The definitions of prespecified criteria for classification of events are listed in the Appendix . A flowchart of the overall adjudication process is shown in the Appendix .

Each case was reviewed independently by 2 physicians blinded to treatment in the phase I review. The case was classified as resolved if the physicians agreed on the mode of death or hospitalization. Cases in which there was disagreement between the 2 physicians were forwarded to a second-level (phase II) review for adjudication by consensus of the full committee. The committee members were blinded to treatment and to the result of the phase I review. If necessary, physicians could request additional medical records to adjudicate a suspected event. If additional records were obtained, the case was reviewed again to ensure that decisions were based on similar documents. For quality assurance, 10% of the cases with agreement by the phase I physicians were reviewed in a blinded fashion by the phase II committee to determine whether there were any systematic inconsistencies with the phase I reviews.

Subject demographics were summarized as numbers and percentages for dichotomous variables or mean ± SD and medians (25th percentile, 75th percentile) for continuous variables. Continuous and ordinal variables were compared using Student’s t tests, and chi-square tests were used to calculate p values for comparisons. Percentage agreement was calculated by dividing the number of events with agreement by the total number of events. Free-marginal κ statistics were calculated to assess the overall rates of agreement. A series of 2 × 2 tables with κ and McNemar tests were conducted for each mode of death and hospitalization. The treatment effect for the trial’s primary end point of time to death or cardiovascular or renal hospitalization at 12 weeks was calculated using Cox proportional-hazards regression modeling for both CEC and CRF end point classification. Data were supplied by Ortho-McNeil Janssen Scientific Affairs (Raritan, New Jersey). Data analysis was performed jointly by the statistical groups at Ortho-McNeil Janssen Scientific Affairs with independent validation at the Duke Clinical Research Institute. Ortho-McNeil Janssen Scientific Affairs had no other role in any aspect of the study.

Results

The baseline characteristics are listed in Table 1 . Most of the clinical characteristics were similar between patients who died and those who survived, with similar age and gender, but those who died had lower baseline left ventricular ejection fractions (23% vs 25%), lower systolic and diastolic blood pressures, lower serum sodium, and higher blood urea nitrogen. Of note, those who died had higher baseline serum creatinine than those who survived (mean 1.8 vs 1.6 mg/dl). There were no important differences in concomitant medications at baseline, except higher use of angiotensin-receptor blockers in patients who survived (16% vs 7%) and higher use of loop diuretics in patients who died (86% vs 76%).

There were 151 deaths in the study at 24 weeks, including 1 patient who did not receive treatment and was not included in the 911 patients used for the efficacy analyses at 12 weeks. Two patients died within a 5-day window of the 24-week end point and were used in generating the comparison of modes of death. There were no statistically significant differences in the risk for death between the nesiritide and placebo groups, as previously reported. In addition, there were no differences in the number of deaths or time to death between the CEC and site investigator reporting, and there was no difference in the primary end point of all-cause death or time to death or cardiovascular or renal hospitalization at 12 weeks using site investigator–determined events versus CEC-adjudicated events.

The distribution of mode of death is shown in Figure 1 . Most deaths were determined to be cardiac (63% by the CEC and 77% by site investigator assessment), followed by noncardiovascular as the second most common mode of death ( Table 2 ). No deaths were classified by either the CEC or site investigators as due only to renal causes.

Mode of death distribution. CV = cardiovascular.

There were 1,041 hospitalization events in the 24-week study follow-up period. The distribution of mode of hospitalization is shown in Figure 2 . As with mode of death, most hospitalizations were determined by site investigators and the CEC to be cardiac related, followed by noncardiovascular, with 1% (CEC) of hospitalizations attributed to pure renal causes and approximately 10% (CEC) that could not be classified.

Reasons for hospitalization distribution. CV = cardiovascular.

The rates of agreement and discordance between investigators and the CEC concerning cause of death on the basis of 2 × 2 tables for each mode are listed in Table 3 . There were differences in agreement between the investigators and the CEC for cardiac or noncardiovascular causes of death (McNemar p <0.001 and p = 0.04, respectively). The overall rate of agreement for all causes of death was 72% (κ = 0.65).

The agreement and discordance on reasons for hospitalization are listed in Table 4 . Overall, there was 67% agreement (κ = 0.59) on all categories of hospitalization. McNemar tests for each of the hospitalization categories using 2 × 2 tables indicated significant differences in agreement between the investigators and the CEC in all categories, with the most disagreement related to cardiac cause of death (25% disagreement). The level of agreement on the basis of κ statistics ranged from poor agreement for cardiorenal and unknown causes to good agreement for noncardiovascular causes (κ = 0.66, McNemar p = 0.001). There was moderate agreement on cardiac causes for hospitalization.

The rates of agreement and discordance concerning mode of death by region of the world are listed in Tables 5 and 6 . There were more noncardiovascular deaths in the rest of the world than in North America (16% vs 5%, respectively), with similar rates of cardiac-related death of 55% and 58%, respectively. There was better agreement between investigators and the CEC in the rest of the world than in North America for mode of death. There were greater differences in agreement for cardiac and cardiorenal modes of death in North America. All modes of death that occurred in North America had fair interrater agreement between investigators and the CEC.

In contrast, none of the modes of death that occurred in the rest of the world had evidence of significant disagreement between investigators and the CEC. The rates of agreement were high at 84% for cardiac and 94% for cardiorenal causes of death. Degree of agreement ranged from fair to very good for the modes of death in the rest of the world.

The rates of agreement for mode of hospitalization also differed by region of the world ( Tables 7 and 8 ). There was no difference in agreement on cardiac reasons for hospitalization with 78% rate of agreement (McNemar p = 0.36) as well as noncardiovascular reasons for hospitalization. The degree of agreement in North America ranged from poor to good. With respect to the rest of the world, there were significant differences in agreement for cardiac reasons for hospitalization, with disagreement on 31% (McNemar p <0.001), as well as noncardiovascular and unknown causes for hospitalization, with disagreement on 17% and 37%, respectively (McNemar p <0.001). The degree of agreement concerning reasons for hospitalization for the rest of the world ranged from poor to fair.

Table 7

Cross table for reasons for rehospitalization: investigators versus clinical events committee by region: North America (n = 774)

Mode	% Agree	% Disagree	κ (95% CI)	McNemar ⁎	Test κ = 0 †
Cardiac	600 (78%)	174 (22%)	0.53 (0.47 to 0.59)	p = 0.36	p <0.0001
Cardiorenal	751 (97%)	23 (3%)	0.37 (0.16 to 0.57)	p = 0.01	p <0.0001
Renal	678 (88%)	96 (12%)	0.10 (0.01 to 0.19)	p <0.0001	p = 0.0004
Other or noncardiovascular	684 (88%)	90 (12%)	0.71 (0.66 to 0.77)	p = 0.83	p <0.0001
Unknown	729 (94%)	45 (6%)	−0.01 (−0.02 to 0.0004)	p <0.0001	p = 0.68

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Stent-Related Cardiac Events Beyond Three Years After Implantation of the Sirolimus-Eluting Stent (from the EVASTENT Patients)
2. A Propensity-Matched Study of the Comparative Effectiveness of Angiotensin Receptor Blockers Versus Angiotensin-Converting Enzyme Inhibitors in Heart Failure Patients Age ≥65 Years
3. Comparison of Ascending Aortic Size in Patients With Severe Bicuspid Aortic Valve Stenosis Treated With Versus Without a Statin Drug
4. Henry Bowditch and “The Young Stethoscopist”

Stay updated, free articles. Join our Telegram channel

Join